Maybe a very basic question but I hope someone can help. Discussion and use of different Canonical ordinarios methods (redundancy analysis RDA, canonical correspondence analysis CCA, linear discrimination análisis LDA, canonical correlation analysis CCorA, partial least squares PLS), to explicitly explore patterns of covariation or relationships between 2 or more data sets, are common in the ecological literature.
However, in the case of geometric morphometrics, the methods most used or suggested to explore relationships between a set of morphological variables and a set of other variables (environmental, ecogeographic, trophic), basically focus on symmetrical Canonical ordinarios methods as "Canonical correlation analysis" and mainly PLS analysis (after Rohlf & Corti, 2000). Perhaps the explanation is very simple and is related to the properties of geometric morphometrics data. But I have severe doubts on the applicability (or not) the asymmetric methods (e.g. Redundancy analysis), which are very common and powerful in ecology, especially in cases in which environmental variables are assumed as explanatory variables and those who do not comply with the assumption of a linear relationship between environmental variables and morphological variables assumed in the PLS. Thanks. LUIS. -- MORPHMET may be accessed via its webpage at http://www.morphometrics.org To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
